Activity Number:
|
580
- Methodological Developments and Implications for Social Scientists
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, July 31, 2019 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Social Statistics Section
|
Abstract #307372
|
Presentation
|
Title:
|
Covariate Selection in Small Randomized Studies
|
Author(s):
|
David Judkins*
|
Companies:
|
Abt Associates, Inc.
|
Keywords:
|
|
Abstract:
|
Considerable progress has been made in recent years in the asymptotic properties of procedures for covariate control in randomized studies. This paper focuses on the special challenges in small randomized studies. Simulation studies demonstrate that a cross-validated LASSO is an excellent choice. No selection at all results in power loss compared to no covariate adjustment at all. Stepwise procedures result lower variances but also in underestimated variances and excess type I error.
|
Authors who are presenting talks have a * after their name.
Back to the full JSM 2019 program
|